Using pip Instead of uv in This Project

This project uses uv in documentation and examples for speed and reproducibility, but you can do everything with pip and standard Python commands.

Installation

What uv sync does is it creates a virtual environment (if not already created) and installs the dependencies listed in pyproject.toml along with the project itself. So, to replicate that with pip, you can do:

# Create a virtual environment (if not already created)
python3.13 -m venv .venv

# Activate the virtual environment
# On Windows
.venv\Scripts\activate
# On Unix or MacOS
source .venv/bin/activate

# Upgrade pip and install dependencies (optional but recommended)
pip install --upgrade pip setuptools wheel

# Install the project and its main dependencies
pip install .

# Optional: Install extra dependencies as needed
# For enhanced CLI features (`prompt-toolkit`)
pip install ".[cli]"
# For tracking and monitoring (`langfuse`)
pip install ".[monitoring]"
# For all optional dependencies
pip install ".[cli,monitoring]"

Running the MCP Server and Client

The package provides two command-line entry points: mcp-server and mcp-client. You can run these commands directly after installation.

# Running the MCP Server
mcp-server --log_level INFO

# Running the MCP Client
mcp-client --language_model "gpt-4.1" azure_openai --azure_openai_endpoint "<YOUR-ENDPOINT-URL>" --azure_openai_deployment_name "gpt-4.1" --azure_openai_api_version "2025-01-01-preview" --azure_openai_api_key "<YOUR-API-KEY>"